Opioid-induced respiratory depression is antagonized effectively by the competitive opioid receptor antagonist naloxone. However, to fully understand the complex opioid agonist-antagonist interaction, the effects of various naloxone doses on morphine and morphine-6-glucuronide (M6G)-induced respiratory depression were studied in healthy volunteers.


Twenty-four subjects received 0.15 mg/kg morphine intravenously at t = 0 followed by placebo, 200 or 400 microg naloxone at t = 30 min. Thirty-two subjects received 0.3 mg/kg M6G intravenously at t = 0 followed by placebo, 25, 100, or 400 microg naloxone at t = 55 min. There were a total of 8 subjects per treatment group. Respiration was measured on a breath-to-breath basis at constant end-tidal Pco2. A mechanism-based pharmacokinetic-pharmacodynamic model consisting of a part describing biophase equilibration and a part describing receptor association-dissociation kinetics was used to analyze the data.


Naloxone reversal of M6G-induced respiratory depression developed more slowly than reversal of the respiratory effect of morphine. A simulation study revealed that this was related to the slower receptor association-dissociation kinetics of M6G (koff M6G = 0.0327 +/- 0.00455 min versus morphine 0.138 +/- 0.0148 min; values are typical +/-SE). Duration of naloxone reversal was longer for M6G. This was related to the three- to fourfold greater potency of naloxone as an antagonist against M6G compared with morphine. Increasing the naloxone dose had no effect on the speed of reversal, but it did extend reversal duration.


Naloxone reversal of the opioid effect is dependent on the receptor association-dissociation kinetics of the opioid that needs reversal with respect to the rate of reversal. The pharmacodynamics of naloxone determines reversal magnitude and duration.

What We Already Know about This Topic

  • ❖ Reversal of respiratory depression from fentanyl and buprenorphine by naloxone can be explained by distribution kinetics into the biophase and ligand kinetics at the receptor

  • ❖ Such an analysis has not been applied to morphine or its active metabolite morphine-6-glucuronide (M6G)

What This Article Tells Us That Is New

  • ❖ In 24 healthy volunteers, naloxone reversed M6G more slowly than morphine, consistent with slower receptor association–dissociation kinetics of M6G

  • ❖ Duration of naloxone was greater for M6G, reflecting a greater potency of naloxone as an antagonist to M6G than morphine

THE μ-opioid receptor is the primary drug target for treatment of severe acute pain. Opioid activation of μ-opioid receptors expressed on neurons involved in the control of breathing (primarily located in the brainstem) will cause diminished breathing activity and consequently increased arterial carbon dioxide concentrations.1 

Opioid agonists are considered relatively safe when administered under clinical protocols and guidelines. A recent review of the literature on opioid-related respiratory events in perioperative patients indicates a “low” figure of 0.5% for respiratory depression requiring immediate action.1Although this is true for the typical perioperative patient, it is evident that there are various patients groups at higher risk for respiratory events related to opioid administration (e.g ., patients with obstructive sleep apnea, the morbidly obese, and the very old).1An important tool to treat opioid-induced respiratory depression is by direct antagonism of the opioid effect using a μ-opioid receptor antagonist. Two opioid antagonists are available clinically as rescue medication for serious opioid-induced side effects, naloxone and naltrexone, but only naloxone is approved for therapeutic use in the reversal of opioid-induced activity and adverse reactions perioperatively, including respiratory depression.2Naloxone is a nonselective competitive opioid antagonist. The ability of naloxone to effectively reverse opioid-induced respiratory depression depends on many factors, including the pharmacologic properties of the opioid that needs reversal and its interaction with naloxone. Therefore, effective and safe reversal of opioid-induced respiratory depression is essential to understand the pharmacokinetics and pharmacodynamics of the specific opioid agonist, the antagonist, and their interactions.1 

Recently, we applied a mechanism-based (i.e ., incorporating concepts from receptor theory) pharmacokinetic–pharmacodynamic (PK-PD) model to study intravenous opioid-induced respiratory depression and its reversal by naloxone in humans, focusing on the partial μ-opioid receptor agonist, buprenorphine, and full agonist, fentanyl.3,4The model had a part that describes biophase distribution kinetics (passage across the blood–brain barrier) and a part that describes receptor association–dissociation kinetics. For buprenorphine, it was shown that slow biophase distribution combined with slow receptor association and dissociation kinetics determines the dynamics of the observed respiratory depression that is characterized by slow onset and long duration of action.3The slow receptor association–dissociation is related to the fact that buprenorphine does not bind immediately to the target receptor but eventually binds with high affinity. In contrast, fentanyl receptor binding occurs rapidly, and, consequently, the short delay in effect (hysteresis) is well explained solely by biophase distribution kinetics.4The slow receptor association–dissociation kinetics of buprenorphine determines the difficulty in producing rapid reversal of buprenorphine-induced respiratory depression by naloxone.4,5In a separate study, it was shown that reversal of buprenorphine-induced respiratory depression is only feasible when it is administered as a continuous infusion of 4–8 mg naloxone/h; however, the rate at which reversal was complete remained slow, taking about 1 h, and was independent of the naloxone dose.6The results of these studies indicate that application of mechanism-based PK-PD models constitutes a clinically relevant basis for the design of naloxone-dosing regimens.

In this study, we applied the mechanisms-based PK-PD model to describe naloxone reversal of morphine- and morphine-6-glucuronide (M6G)-induced respiratory depression. Our aim was to characterize the in vivo  pharmacodynamic interaction of morphine and M6G with naloxone, further extending our knowledge of naloxone reversal of opioid-induced respiratory depression. We chose to study morphine as it is considered as the definitive standard of opioid treatment. M6G is an active metabolite of morphine and, as such, may be important in contributing to morphine-induced respiratory depression, especially when morphine treatment is prolonged or when renal excretion of the metabolite is compromised. Taking into account the recent data on the interaction between naloxone and buprenorphine (cf . Refs. 3, 4, and 6), we hypothesize that (1) naloxone readily reverses respiratory depression induced by morphine and its active metabolite M6G, that is, at naloxone doses less than 800 μg; (2) the rate (speed) of reversal is dictated by the receptor kinetics of the opioid agonist, that is, predominantly dependent on the receptor dissociation kinetics; and (3) that the opioid agonist with the slowest receptor dissociation kinetics will show the slowest reversal by naloxone. Because there are no data on the receptor association–dissociation kinetics of opioid receptors involved in respiratory depression for morphine and M6G, we cannot predict at this point which of the two opioids is reversed faster or slower.

We performed studies in healthy volunteers and obtained high-resolution (i.e ., breath to breath) ventilation data obtained at clamped end-tidal oxygen and carbon dioxide concentrations and tested various naloxone doses. This approach allows optimal conditions for studying drug interactions on respiratory depression.3,4,6,7 

Materials and Methods


Fifty-six male and female volunteers (age range, 18–34 yr, body mass index < 28) were recruited to participate in the studies after obtaining approval of the protocol from the Leiden University Medical Center Human Ethics Committee (Commissie Medische Ethiek, Leiden, The Netherlands) and after giving written informed consent. The trial was registered at #NTR237 (#ISRCTN 59442355). All candidates underwent a physical examination, and only healthy subjects without a history of illicit drug use or psychiatric illness were allowed in the study. All women used oral contraceptives. All subjects were advised not to eat or drink for at least 8 h before the start of the study.


After arrival to the laboratory, the subjects were seated comfortably on a hospital bed and received two intravenous lines for drug infusion (one for morphine or M6G and one for naloxone or placebo). During the studies, subjects breathed through a facemask (fitted over nose and mouth). The subjects were asked to breath through their mouth to minimize changes in airway resistance. The mask allows normal movement of mouth, lips, and tongue, and it is considered less disruptive to normal breathing than a mouthpiece. The airway gas flow was measured with a pneumotachograph (#4813; Hans Rudolph, Myandotta, MI) connected to a pressure transducer that yields a volume signal. The pneumotachograph was heated throughout the 240-min study period. This signal was calibrated with a 1 l calibration syringe (Hans Rudolph). The pneumotachograph was connected to a T-piece; one arm of the T-piece received a gas mixture with a flow of 45 l/min from a gas mixing system, consisting of three mass-flow controllers (Bronkhorst High-Tec, Veenendaal, The Netherlands) through which the flow of oxygen, nitrogen, and carbon dioxide could be set individually at any desired level. A computer provided control signals to the mass-flow controllers, allowing adjustment of the inspired gas mixture to force the end-tidal gas concentrations of oxygen and carbon dioxide to follow a specific pattern in time (end-tidal forcing).8,9In this study, the end-tidal oxygen concentration was maintained at 110 mmHg, whereas the end-tidal carbon dioxide concentration was maintained such that ventilation was increased to 25 ± 2 l/min.3,4,6,7For each subject, the end-tidal carbon dioxide concentration was titrated to effect to reach this ventilation value. In this study, the end-tidal carbon dioxide concentration ranged between 46 and 49 mmHg. Gas concentrations were measured with a gas analyzer (Datex Multicap, Helsinki, Finland); arterial hemoglobin oxygen saturation was measured via  a finger probe (Spo2) with a Massimo pulse oximeter (Irvine, CA). The software programs for steering end-tidal gas concentrations and data acquisition (ACQ and RRDP) were custom built (Erik Kruyt, M.Sc., and E. Olofsen, LUMC, Leiden, The Netherlands). End-tidal oxygen concentration, end-tidal carbon dioxide concentration, minute ventilation (V̇E), and Spo2were collected on a breath-to-breath basis and stored on a disc for analysis.

Study Design

The study had a single-blinded, placebo-controlled, and randomized design.

Study Morphine.

Twenty-four subjects received intravenous morphine (0.15 mg/kg) at time t = 0, followed by placebo (normal saline, n = 8), 200 (n = 8), or 400 μg (n = 8) of intravenous naloxone at t = 30 min. Breathing was measured from 5 min before morphine infusion until 120 min continuously.

Study M6G.

Thirty-two subjects received intravenous M6G (0.3 mg/kg) at t = 0, followed by placebo (normal saline, n = 8), 25 (n = 8), 100 (n = 8), or 400 μg (n = 8) intravenous naloxone at t = 55 min. Breathing was measured for 5 min before M6G infusion and from t = 55 min for at least 65 min (but no longer than 100 min), followed by 5-min measurements at t = 150, 180, 210, and 240 min. The continuous measurement of ventilation started just before the naloxone infusion. Attempts were made to measure ventilation continuously from t =−5 to 150 min but were abandoned because of discomfort of the subjects while breathing via  the facemask for such a long time (> 2.5 h). The current design took this discomfort away. The M6G dose chosen is equivalent in potency to 0.15 mg/kg morphine.10–13 

Administration of naloxone at 30 min after morphine infusion and 55 min after M6G infusion was chosen, as the speed of onset of morphine is faster than that of M6G by about a factor of 2.10–13Our current approach allows infusion of the opioid antagonist at approximately equal levels of respiratory depression from the two opioids.


M6G was donated by CeNeS Ltd. (Cambridge, United Kingdom; now PAION UK Ltd.). Morphine was purchased from Pharmachemie BV (Haarlem, The Netherlands), and naloxone from Orpha-Devel GmbH (Pukersdorf, Austria). Placebo (NaCl 0.9%) was manufactured by the local pharmacy. A physician not involved in the study performed randomization (using computer-generated randomization lists) and preparation of the syringes. All drugs were infused for more than 90 s.

Data Analysis: Descriptive Analysis

The breath-to-breath data were averaged more than 1-min periods. Baseline ventilation (V0), time to maximum reversal (i.e ., the time at which ventilation reached its highest point after infusion of naloxone/placebo; Tmax), and ventilation at its highest point after infusion of naloxone/placebo (V̇MAX) were computed. These “summary measures” were compared within studies for morphine and M6G (analysis of variance, factor = dose). Next, ventilation just before naloxone or placebo infusion (V̇), TMAX, and V̇MAXwere compared between studies (analysis of variance, factors = treatment, dose). Statistical analysis was performed using the statistical package R (version 2.8; R Project).#A value of P  less than 0.05 was considered significant. Values given are mean ± SEM unless otherwise stated.

Data Analysis: Population PK-PD Analysis

In a pilot study, we observed that blood sampling had stimulatory effects on breathing, causing the inability to discern true naloxone effects. Therefore, we decided to perform this study without the drawing of blood. Under these conditions, to be able to perform a PK-PD analysis, we assumed that morphine, M6G, and naloxone concentrations are well described by earlier established pharmacokinetic models (see Refs. 3,10–12). Our current approach is similar to that applied in earlier studies in our laboratory.13Furthermore, most of the subjects participating in this study also participated in these pharmacokinetic studies on morphine, M6G, or naloxone.3,11,12A reanalysis was performed to obtain pharmacokinetic parameters scaled with respect to 70 kg. In this way, pharmacokinetic interindividual variability due to weight is taken into account. Three- (for morphine and M6G) and two- (for naloxone) compartmental models were used with lognormal distributions for the pharmacokinetic parameters across the population and constant relative residual error. Models were parameterized using volumes and clearances or volumes and k values depending on the minimum values of the objective function of NONMEM (see table 1for pharmacokinetic parameter values).

The Pharmacodynamic Model.

The differential equations that describe the opioid agonist (M, i.e ., morphine or M6G) and antagonist (N, i.e ., naloxone) molecules binding to the opioid receptor (R) are:

where konand koffare receptor equilibration rate constants for ligand receptor association and dissociation, respectively.

For naloxone, we assume that koff,Nis large.14We may then assume that


with C50,N= koff,N/kon,N.

After normalizing [MR] and [NR] by setting [R]+[MR]+[NR]= 1, we have:

For an adequate description of the data, we introduce a steepness parameter γ for naloxone:

Biophase distribution kinetics was characterized by parameter t½ke0for each of the three tested drugs. Finally, ventilation (V̇E) was assumed to depend on [MR]:

where V̇0is predrug baseline ventilation. Because [MR] was normalized (i.e ., 0 < [MR] < 1) it follows that 0 < V̇E< V̇0.

Statistical Analysis.

The data were analyzed with the statistical package NONMEM VI version 1.2 (ICON Development Solutions, Ellicott City, MD).15Model parameters were assumed to be log-normally distributed except parameter γ, which was assumed to lie between 0 and 20 via  the inverse logit transformation to avoid numerical problems. Residual error was assumed to be additive with variance ς2. The P  values less than 0.01 were considered significant.

Simulation Study.

To assess the effect of variations in the different pharmacodynamic model parameters on naloxone reversal of opioid-induced respiratory depression, we performed a simulation study using equations 1, 3, and 4. The effect of 100 μg of naloxone given at t = 55 min on top of 21 mg of M6G given at t = 0 min was simulated for a subject weighing 70 kg with variations in parameters M6G t½ke0, M6G kon, M6G koff, and the combination of M6G konand M6G koffsuch that their ratio remains constant, and naloxone t½ke0, naloxone C50, and parameter γ. The parameter values simulated ranged from 20 to 500% of the typical population value estimated in this study (actual values tested 20, 50, 100, 200, and 500%). In the simulations, only one (or two for the combination konand koff) parameter was varied. The other parameter values were fixed to their typical value.


The characteristics of the subjects are given in table 2. All 56 subjects completed the study without major side effects. Minor side effects were similar to those observed in previous studies.10–13Subjects experienced a heavy feeling on chest or in the extremities (lasting 2–5 min) only on infusion of M6G. Nausea (but no vomiting) and sedation were experienced only on infusion of morphine. All symptoms were mild and did not require treatment. None of the subjects fell asleep during the studies, and none snored. Sedation scores (on a scale from 0 to 10 in which 0 is alert and 10 is unable to keep the eyes open) ranged from 2 to 5 for subjects receiving morphine and 0 to 3 for subjects receiving M6G.

The control of end-tidal gas concentrations was good throughout the studies with an average end-tidal oxygen concentration of 110 ± 2 mmHg and an end-tidal carbon dioxide concentration of 48 ± 0.6 mmHg. The inspired carbon dioxide values required to maintain constant end-tidal values were 3% (range, 2–4%) at baseline (predrug values), 1% (range, 0.5–2%) at peak drug effect, and 3% (range, 2–4%) at full reversal. Spo2values remained in the normoxic range (>95%) throughout the studies. Baseline ventilatory variables in the morphine study were as follows: ventilation = 24.8 ± 2.1 l/min, tidal volume = 1.3 ± 0.3 l, and respiratory rate = 19.1 ± 4.4 min−1; values in the M6G group: ventilation = 23.6 ± 1.8 l/min, tidal volume = 1.2 ± 1.0 l, and respiratory rate = 19.7 ± 3.1 l/min. (Values are mean ± SD.)

Descriptive Analysis

Ensemble averages for the effect of placebo and naloxone on morphine and M6G-induced respiratory depression are shown in figures 1 and 2. The dotted lines in the M6G plots (fig. 2) are added to connect the measurement periods because no data were collected between baseline ventilation and ventilation levels at t = 55 min. There was no difference in the magnitude of respiratory depression between morphine and M6G studies: Mean ventilation just before the infusion of placebo or naloxone (V̇) was 0.69 ± 0.03 in the morphine studies relative to baseline versus  0.67 ± 0.01 in the M6G studies (table 3, P > 0.05). For both morphine and M6G, the reversal effect of naloxone was greater than that of placebo as determined by V̇MAX(P < 0.01; figs. 1 and 2and table 3). Within studies, no differences in V̇MAXand TMAXwere observed for different naloxone doses. The effect of naloxone on V̇MAXdid not differ between studies (table 3). In contrast, TMAXdid differ significantly between morphine and M6G studies, by about 30 min, TMAXvalues were 18.3 ± 2.1 min for morphine and 49.3 ± 7.6 min for M6G (P < 0.001). Duration of naloxone reversal was greatest for the highest naloxone dose tested (fig. 2D). The effect of morphine, M6G, and naloxone was proportionally distributed among tidal volume and respiratory rate (data not given).

Population Pharmacodynamic Analysis

The model adequately described the reversal of morphine- and M6G-induced respiratory depression. Examples of best, median, and worst data fits for the two studies are shown in figure 3. Goodness-of-fit plots, showing the individual predicted versus  measured data, for morphine and M6G, are given in figure 4. Model parameter estimates are given in table 4. Significant differences were observed between the two studies. The biophase equilibration half-lives (t½ke0) for morphine and M6G differed by a factor of 2, with morphine having faster equilibration (morphine t½ke0= 1.2 h vs . M6G = 2.7 h). The half-lives for receptor dissociation (ln2/koff= t½koff) also differed with a value of 21 min for M6G and 5 min for morphine. The apparent potency of morphine for depression of ventilation as determined by the ratio koff/kon(KD) was 5.5 times greater than that of M6G: morphine KD= 160 nm versus  M6G KD= 880 nm. Naloxone effect differed between studies with a threefold greater naloxone potency when administered to antagonize M6G compared with morphine (naloxone C50= 0.5 and 1.8 nm in M6G and morphine studies, respectively). The naloxone onset/offset times did not differ between studies, possibly related to the high variability in the values observed in the morphine studies (coefficient of variation = 25%). Finally, the shape parameter γ differed between studies with a steeper value in the M6G study (7.4) than in the morphine study (4.2).

Simulation Study

The results of the simulation study are shown in figures 5 and 6. In figure 4, M6G-related pharmacodynamic parameters are varied between 20 and 500% of their typical population value. Variations in parameter t½ke0has an effect on the magnitude of respiratory depression observed in the first 55 min, but little effect on the speed of naloxone reversal was observed (fig. 5A). In contrast, parameters konand koff(apart from an effect on M6G potency) have an effect on the rate at which naloxone reversal develops (fig. 5C). At koffand konvalues of 20% of the typical population estimate (t½koff> 100 min), the slope of ventilation after naloxone infusion is most shallow, whereas at a value of 500% of typical (t½koff< 5 min), the slope is steepest and full reversal is observed lasting for about 30 min. A similar picture emerges when viewing the effect of changing both konand koffwith the constraint that their ratio remains constant at the estimated KDvalue of 880 nm. Only at large values of koffand kon(yellow line) reversal is rapid and complete (i.e ., ventilation reaches predrug baseline values). In figure 6, naloxone-related pharmacodynamic parameters are varied. It is obvious that naloxone pharmacodynamic parameters have no effect on the rate at which reversal of respiratory depression develops. Variations in parameter t½ke0have little effect on reversal. Variations in naloxone potency have an effect on the magnitude of reversal and duration (fig. 6A): the more potent naloxone is the greater is the magnitude of reversal and its duration (fig. 6B). Finally, shape parameter γ has an effect on the magnitude of reversal although the effect is rather limited (fig. 6C).


We previously modeled the effect of morphine and M6G on breathing and pain relief and measured arterial plasma concentration as an input to the PK-PD models.10–12In contrast, in this study, we analyzed the data using simulated pharmacokinetic data based on earlier established pharmacokinetic models from our laboratory. In this study, we refrained from obtaining plasma samples as we observed in a pilot study that frequent sampling had excitatory effects on breathing. This rendered us unable to describe the reversal effect of naloxone accurately because blood sampling occurring around the time of naloxone infusion exaggerated the effect of naloxone. Simulated pharmacokinetic data have been applied successfully in previous PK-PD studies when we modeled the effect of morphine and M6G on normoxic and hypoxic breathing.13Although we agree that the lack of pharmacokinetic data is a possible drawback of our study, we believe that our approach is valid as the variance in the pharmcokinetic datasets was minimal (table 1), and most importantly, the majority of the current volunteers had participated in these previous pharmacokinetic studies.3,11,12Furthermore, in contrast to previous studies,10–12the collected breath-to-breath data were used as pharmacodynamic model output. These high-resolution datasets may have increased our ability to obtain reliable parameter estimates.

The mechanism-based PK-PD model that we applied enabled the estimation of parameters related to biophase distribution and receptor association–dissociation kinetics. As such, the kinetics of drug action is governed by rate constants ke0, kon, and koff. Our in vivo  morphine and M6G pharmacodynamic parameter values are in good agreement with earlier in vivo  and in vitro  observations. For example, the estimated KD(koff/kon) values or apparent potencies are of the same order of magnitude as the morphine and M6G C50values observed in human respiratory and analgesia studies.10–13Similarly, in agreement with earlier findings,10–12the biophase distribution half-life of M6G, t½ke0, was slower than that of morphine by a factor of 2. Most in vitro  studies indicate that M6G has more than fourfold lower affinity (as assessed by radioligand binding) for the μ-opioid receptor than morphine,16which compares well with the 5.5-fold difference observed in KDin this study. These observations suggest that our analysis yielded reliable parameter estimates and hence give sufficient validation to our modeling approach. Evidently, further studies using plasma sampling will be needed to confirm our results.

For naloxone, we assumed rapid association/dissociation kinetics. The in vitro  half-time for dissociation of the naloxone from the μ-opioid receptor complex is less than 1 min, using CHO-K1 cells transfected with the human μ-opioid receptor.14We did observe values for t½ke0ranging from 5.4 to 11.2 min, which are in agreement with earlier findings on the naloxone-buprenorphine interaction with a range for naloxone t½ke0of 5 to 8 min3. The elimination half-life of naloxone from plasma is 33 min3. Because the distribution of naloxone in and out of the brain compartment and receptor kinetics are faster than its (rapid) elimination half-life from plasma, the latter is the major limiting factor in determining the duration of the presence of naloxone at its target site. This is exemplified by the naloxone–buprenorphine interaction by which buprenorphine exhibits a slow biophase distribution, which, together with slow receptor association and dissociation kinetics, determines that buprenorphine-induced respiratory depression is characterized by slow onset and long duration of action.3–6Under these circumstances, a bolus infusion of 800 μg naloxone has no effect on buprenorphine-induced respiratory depression as plasma levels fall rapidly and naloxone is washed out from the brain compartment before any significant reversal of burprenorphine sets in. Only with a continuous infusion and hence the continuous presence of naloxone at its target site does any reversal occur albeit at a slow pace. In this study, the reversal of morphine- and M6G-induced respiratory depression by a bolus naloxone infusion was possible, although there was a significant difference in reversal speed and duration of action.

The three- to fourfold difference in naloxone potency (C50) observed in morphine and M6G studies was unexpected. A possible explanation for the higher naloxone potency against M6G is that M6G may increase naloxone affinity for the target receptor. Abbott and Palmour17described such a phenomenon when testing the binding of radiolabeled opiates to rat brain membrane receptors. At low concentrations, M6G but not morphine enhanced the binding of [H3]naloxone to rat brain receptors by 20–40%. An alternative explanation is that morphine and M6G act at different receptors within the brain compartment each with a distinct naloxone affinity. A unique M6G receptor has been postulated previously,18but we were unable to demonstrate its existence in an animal model of M6G versus  morphine-induced respiratory depression.19Shafer et al .20postulated that if a drug would act at several targets in the signaling cascade, this would increase the apparent potency as well as the value of the steepness parameter γ. We added a steepness parameter to describe the data adequetly. If naloxone would act at various sites within the signaling cascade, this could explain the need for parameter γ (>1), and when these sites would differ for morphine and M6G, this could explain the observed differences in C50values. Possible target sites, apart from the μ-opioid receptor, include a second target on the μ-opioid receptor, an influence on the G-protein opioid-receptor complex, neuronal K+-ion channels, or a second (M6G unique) opioid receptor.

Naloxone reversal of M6G-induced respiratory depression was slower than that of morphine by about 30 min (table 3), but with 400 μg naloxone the reversal effect lasted longer (figs. 1 and 2). The results of our PK-PD analysis and subsequent simulation study indicate that the rate (or speed) of reversal is dependent on the receptor kinetics of the opioid that needs reversal. This is in agreement with our previous findings for buprenorphine, which has lower k-values than M6G (kon= 0.0096 nm.min−1and koff= 0.017 min−1) and cannot be reversed by a bolus naloxone infusion.3,4,6Combining the previous and current results, it follows that an opioid with lower values for konand koffis more difficult to reverse with naloxone, and consequently reversal develops more slowly or requires a continuous infusion (see figs. 5B–D). The order of difficulty for naloxone to reverse the three opioids tested by us so far, from most difficult to easiest, is as follows: buprenorphine > M6G > morphine. Although we did not test the naloxone–fentanyl interaction, the receptor kinetic data estimated from fentanyl respiratory studies indicate that it would be even easier than morphine to reverse with naloxone (fentanyl kon> 100 nm/min and koff> 100 min−1),4that is, buprenorphine > M6G > morphine > fentanyl. In contrast, the magnitude and duration of reversal depend to a greater extent on the pharmacodynamic parameters of naloxone than on pharmacodynamic parameters related to the opioid agonist (figs. 5 and 6). The difference in naloxone potency in morphine and M6G studies is the major cause for the difference in the duration of 400 μg naloxone-induced reversal between the two studies (compare fig. 1C and 2D). This is confirmed by the simulation study (see fig. 6B) that shows an increase in duration and magnitude of reversal with increasing values for naloxone C50. However, the receptor kinetics of the opioid agonist does have some influence on the magnitude of naloxone reversal (but not on the duration of reversal). At high values of konand koff(at 200 and 500% of the estimated values given in table 4) of the agonist, we did observe an increase in the magnitude of reversal (fig. 5D). Full reversal of the morphine effect was observed at 200 μg naloxone, whereas 100 μg was sufficient for M6G. Increasing the naloxone dose has a similar effect to increasing its potency, indicating that the reversal of the morphine effect may be extended by giving doses greater than 200 μg (morphine) and 100 μg (M6G) (see figs. 2B–D).

It is important to understand that the inability to increase the speed of reversal of morphine- and M6G-induced respiratory depression by using higher doses of naloxone is due to the fact that reversal is predominantly dictated by receptor association–dissociation kinetics of the opioid agonist and not to that of the opioid antagonist. When receptor kinetics of the agonist is fast (e.g ., as is the case for fentanyl), it does not play a pivotal role in reversal kinetics, and higher doses of the antagonist will cause a speedier reversal. This is, for example, also the case for reversal of muscle relaxation from vecuronium by neostigmine. However, when the receptor kinetics of the agonist is slow (as here in the case of morphine or M6G), the pharmacokinetics and pharmacodynamics of the agonist now determine the speed of reversal. Our data indicate that while both morphine and M6G may be reversed by naloxone, reversal of the M6G effect by naloxone will be slower than that for morphine. Furthermore, giving greater doses of naloxone will have no effect on the speed of reversal, although it will extend the duration of action and the magnitude of effect.

Previously, we showed important sex differences in the effect of morphine, but not M6G, on ventilatory control and analgesia.10,11,21,22In line with our previous findings on morphine, we cannot exclude a possible sex-dependent effect of naloxone (or some influence of the use of oral contraceptives) on reversal of respiratory depression. However, our study was not designed or powered to study possible gender (or hormonal influences) on naloxone effect. Further studies are needed to investigate these important issues. We made use of the computer-controlled dynamic–end-tidal forcing technique.8,9The most important aspect of this technique is the manipulation of inspired carbon dioxide concentrations to maintain a constant end-tidal carbon dioxide concentration throughout the experiment. The technique has been applied in many studies to study physiology and investigate the effect of drugs on the control of breathing in humans and animals.8,9,23,24The technique is a tool to study physiology and pharmacology at the best possible experimental conditions (e.g ., by eliminating the confounding effect of fluctuations in end-tidal carbon dioxide concentration).25However, the experimental conditions that we created are not a full representation of the perioperative patient. Breathing in the perioperative patient is under the influence of many factors, such as respiratory drive, arousal state, and the functionality of the pharyngeal dilating muscles. Opioids do have an effect on all three. In this study, we explored the effect of morphine and M6G and their interaction with naloxone on just the ventilatory drive. The effect of naloxone on the changes in arousal state and upper airway potency from opioids needs further study.

In conclusion, we applied a mechanism-based competitive agonist–antagonist interaction model that describes naloxone reversal of morphine- and M6G-induced respiratory depression with success in human volunteers. Full naloxone reversal was possible for the two opioids tested at doses of 200 μg for morphine and 100 μg for M6G. The rate (speed) of naloxone reversal was dependent on the opioid agonist receptor association–dissociation kinetics: lower values of konand koffmake reversal more difficult, and consequently the process develops more slowly. Hence, reversal was slowest for M6G as it had smaller k values compared with morphine. The potency of naloxone determines the duration and magnitude of reversal with an increase in potency linked to an increase in the reversal duration and magnitude. Because naloxone potency was greater in the M6G study (by factor 3–4), reversal duration with 400 μg naloxone was greater for M6G than for morphine. An increase in naloxone dose will increase the duration of reversal but the speed of reversal will remain unaffected. This has evident clinical implications for postoperative care.


Dahan A, Aarts L, Smith TW: Incidence, reversal and prevention of opioid-induced respiratory depression (Review). Anesthesiology 2010; 112:226–38
Goodman AJ, Le Bourdonne B, Dolle RE: Mu opioid receptor antagonists: Recent developments. ChemMedChem 2007; 2:1552–70
Yassen A, Olofsen E, van Dorp E, Sarton E, Teppema L, Danhof M, Dahan A: Mechanism-based pharmacokinetic–pharmacodynamic modeling of the reversal of buprenorphine-induced respiratory depression by naloxone. Clin Pharmacokinet 2007; 46:966–80
Yassen A, Olofsen E, Romberg R, Sarton E, Teppema L, Danhof M, Dahan A: Mechanism-based PK/PD modeling of the respiratory depressant effect of buprenorphine and fentanyl in healthy volunteers. Clin Pharmacol Ther 2007; 81:50–8
Gal TJ: Naloxone reversal of buprenorphine-induced respiratory depression. Clin Pharmacol Ther 1989; 45:66–71
van Dorp E, Yassen A, Sarton E, Romberg R, Olofsen E, Teppema L, Danhof M, Dahan A: Naloxone reversal of buprenorphine-induced respiratory depression. Anesthesiology 2006; 105:51–7
Nieuwenhuijs D, Olofsen E, Romberg RR, Sarton E, Ward D, Engbers F, Vuyk J, Mooren R, Teppema LJ, Dahan A: Response surface modeling of remifentanil–propofol interaction on cardiorespiratory control and bispectral index. Anesthesiology 2003; 98:312–22
Dahan A, deGoede J, Berkenboch A, Olievier I: The influence of oxygen on the ventilatory response to carbon dioxide in man. J Physiol (Lond) 1990; 428:485–99
Dahan A, Nieuwenhuijs D, Teppema L: Plasticity of central chemoreceptors: Effect of bilateral carotid body resection on central CO2sensitivity. PLoS Med 2007; 4:e239
Sarton E, Olofsen E, Romberg R, den Hartigh J, Kest B, Nieuwenhuijs D, Burm A, Teppema L, Dahan A: Sex differences in morphine analgesia: An experimental study in healthy volunteers. Anesthesiology 2000; 95:1245–54
Romberg R, Olofsen E, Sarton E, den Hartigh J, Taschner PEM, Dahan A: Pharmacokinetic–pharmacodynamic modeling of morphine-6-glucuronide–induced analgesia in healthy volunteers. Anesthesiology 2004; 100:120–33
Dahan A, Romberg R, Teppema L, Sarton E, Bijl H, Olofsen E: Simultaneous measurement and integrated analysis of analgesia and respiration after an intravenous morphine infusion. Anesthesiology 2004; 101:1201–9
Romberg R, Olofsen E, Sarton E, Teppema L, Dahan A: Pharmacodynamic effect of morphine-6-glucuronide versus  morphine on hypoxic and hypercapnic breathing in healthy volunteers. Anesthesiology 2003; 99:788–98
Cassel JA, Daubert JD, DeHaven RN: Alvimopan binding to the μ-opioid receptor: Comparative binding kinetics of opioid antagonists. Eur J Pharmacol 2005; 520:29–36
Beal BL, Sheiner LB, Boeckman AJ: NONMEM User's Guide. Ellicott City, Maryland, ICON Development Solutions, 1989–2006
Ellicott City, Maryland
ICON Development Solutions
Kilpatrick GJ, Smith TW: Morphine-6-glucuronide: Actions and mechanisms. Med Res Rev 2005; 25:521–44
Abbott FV, Palmour RM: Morphine-6-glucuronide: Analgesic effects and receptor binding profile in rats. Life Sci 1988; 43:1685–95
Schuller A, King M, Zhang J, Bolan E, Pain Y, Morgan DJ, Chang A, Czick ME, Unterwald EM, Pasternak GW, Pintar JE: Retention of heroin and morphine-6β-glucuronide analgesia in a new line of mice lacking exon 1 of MOR-1. Nat Neurosci 199; 2:151–6
Teppema LJ, van Dorp ELA, Mousavi Gourabi B, van Kleef JW, Dahan A: Differential effect of morphine and morphine-6-glucuronide on the control of breathing in the anesthetized cat. Anesthesiology 2008; 109:689–97
Shafer SL, Hendrickx JFA, Flood P, Sonner J, Eger EI II: Additivity versus  synergy: A theoretical analysis of implications for anesthetic mechanisms. Anesth Analg 2008; 107:507–24
Dahan A, Sarton E, Teppema L, Olievier C: Sex-related differences in the influence of morphine on ventilatory control. Anesthesiology 1998; 88:903–13
Sarton E, Teppema L, Dahan A: Sex differences in morphine-induced ventilatory depression reside within the peripheral chemoreflex loop. Anesthesiology 1999; 90:1329–38
van de Elsen M, Dahan A, DeGoede J, Berkenbosch A, van Kleef J: Influences of subanesthetic isoflurane on ventilatory control in humans. Anesthesiology 1995; 83:478–90
Berkenbosch A, Teppema LJ, Olievier CN, Dahan A: Influences of morphine on the ventilatory response to isocapnic hypoxia. Anesthesiology 1997; 86:1342–9
Olofsen E, Boom M, Nieuwenhuijs D, Sarton E, Teppema L, Aarts L, Dahan A: Modeling the non–steady state respiratory effects of remifentanil in awake and propofol-sedated healthy volunteers. Anesthesiology 2010; 112:1382–95